Mathematics and Computational Sciences

Revolutionizing Car Insurance Claims

Document Type : Original Article

Authors

Monica Selvam 1
Mohitha Rajesh 1
Sumathi Sundaramoorthy 2

1 St. Josephs Institute of Technology, Chennai, India

2 St. Josephs College of Engineering, Chennai, India

https://doi.org/10.30511/mcs.2026.2077829.1627
Abstract
Abstract-Assessing car damage for insurance claims has traditionally been a slow and often inconsistent process. Typically, it entails manual inspections, which can be laborious, arbitrary, and susceptible to fraud or mistakes. Many insurers are increasingly using artificial intelligence, particularly deep learning and computer vision, to address these issues and make processes quicker and more dependable. Artificial intelligence (AI) systems can now examine images of damaged cars and determine the kind and degree of damage by utilising strong models like Convolutional Neural Networks (CNNs)., and even pinpoint its location all within seconds. This shift not only speeds up the claims process dramatically, but also reduces costs and removes a lot of guesswork from the equation. In this report, we lay out a full end-to-end solution: from collecting the right kind of image data, training and testing AI models, to integrating them into real insurance workflows. The results show major improvements in how fast and accurately claims are processed, and customers are noticing the difference. In short, AI isn’t just improving damage detection it’s reshaping the entire insurance claims experience to be quicker, more transparent, and fairer for everyone involved.

Keywords

Keyword: Deep Learning
Convolutional Neural Networks
Image-Based Analysis
Damage Localization
Claims Processing Efficiency
Subjects
[1] Bochkovskiy, A., Wang, C. Y., & Liao, H. Y. M. (2020). YOLOv4: Optimal speed and accuracy of object detection. arXiv preprint arXiv:2004.10934. 
[2] Dosovitskiy, A., Beyer, L., Kolesnikov, A., et al. (2021). An image is worth 16x16 words: Transformers for image recognition at scale. In International Conference on Learning Representations (ICLR). 
[3] Everingham, M., Van Gool, L., Williams, C. K. I., Winn, J., and Zisserman, A. (2010). The Pascal Visual Object Classes Challenge. International Journal of Computer Vision, 88(2), 303–338. 
[4] He, K., Gkioxari, G., Dollar, P., & Girshick, R. (2020). Mask R-CNN. IEEE Transactions on Pattern Analysis and Machine Intelligence, 42(2), 386–397.
[5] Howard, A. G., Zhu, M., Chen, B., et al. (2017). MobileNets: Efficient convolutional neural networks for mobile vision applications. arXiv preprint arXiv:1704.04861.
[6] Kaggle. (2025). Carvana Image Masking Challenge Dataset. Available at: https://www.kaggle.com/c/ carvana-image-masking-challenge.
[7] Lin, T. Y., Dollar, P., Girshick, R., He, K., Hariharan, B., and Belongie, S. (2017). Feature Pyramid Networks for Object Detection. CVPR.
[8] Lin, T. Y., Goyal, P., Girshick, R., He, K., and Dollar, P. (2017). Focal Loss for Dense Object Detection. ICCV.
[9] Lundberg, S. M., and Lee, S. I. (2017). A Unified Approach to Interpreting Model Predictions. In NeurIPS.
[10] Redmon, J., & Farhadi, A. (2018). YOLOv3: An incremental improvement. arXiv preprint arXiv:1804.02767.
[11] Ren, S., He, K., Girshick, R., & Sun, J. (2017). Faster R-CNN: Towards real-time object detection with region proposal networks. IEEE Transactions on Pattern Analysis and Machine Intelligence, 39(6), 1137–1149.
[12] Samek, W., Wiegand, T., & Muller, K. R. (2019). Explainable artificial intelligence: Understanding, visualizing and interpreting deep learning models. IEEE Signal Processing Magazine, 36(6), 56–65.
[13] Selvaraju, R. R., Cogswell, M., Das, A., et al. (2020). Grad-CAM: Visual explanations from deep networks via gradient-based localization. International Journal of Computer Vision, 128(2), 336–359. 
[14] Tan, M., & Le, Q. V. (2019). EfficientNet: Rethinking model scaling for convolutional neural networks. In Proceedings of the International Conference on Machine Learning (ICML).
[15] The Chinese University of Hong Kong. (2025). CompCars Dataset. Available at: http://mmlab.ie.cuhk.edu.hk/datasets/comp_cars.
[16] Ultralytics. YOLOv8 Documentation. GitHub Repository. Available at: https://github.com/ultralytics/ultralytics.
Mathematics and Computational Sciences
Volume 7, Issue 3
Summer 2026
Pages 117-129

  • Receive Date 14 November 2025
  • Revise Date 19 June 2026
  • Accept Date 19 June 2026

Home

Submit Manuscript

Reviewers

Contact Us